ISA SP100 keeps on its wireless path

ISA’s SP100 Industrial Wireless standardization committee continues its heavy schedule of teleconferences, email debates and frequent face-to-face meetings. The process control suppliers, meanwhile, continue to introduce wireless transmitters and system products at a rapid pace without any claims of standardization. End-user demand for wireless products is rapidly increasing, but end-user participation in the work of SP100 remains sparse.

In September 2006, at the invitation of SP100, suppliers and other interested parties submitted presentations and supporting white papers proposing their wireless solutions—in-whole or in-part. At the time there was a separation between simple data acquisition applications (SP100.14) and more complex systemwide use of wireless for control networks (SP100.11). Some 25 presentations were submitted, along with 15 white papers. When the issues settled, members of the SP100 committee realized that there were few technical differences between the submitted proposals! In the general meeting following the ISA Expo in Houston, the committee unanimously agreed to end the division between the two types of applications and merge the effort into a single wireless solution for the industrial plant and shop floor.

The standard being developed is now labeled SP100.11a, which will define the OSI layer specifications (e.g. PHY, DLL, etc), security specifications, and management (including network and device configuration) specifications for wireless devices serving application classes 1 (closed loop control) through 5 (data acquisition) and, optionally, class 0 (safety) for fixed, portable and moving devices. The standard is to focus on performance needs for periodic monitoring and process control where latencies on the order of 100 ms can be tolerated with optional behavior for shorter latency.

The standard is to provide consideration for the following:

• Low-energy consumption devices
• Scalability to address both small and large installations
• Wireless infrastructure
• Interfaces to and interoperability with legacy infrastructure and applications 
• Security and network management requirements in a functionally scalable manner
• Robustness in the presence of interference found in harsh industrial environments
• Coexistence with other wireless devices anticipated in the industrial work space, such as IEEE 802.11x (Wi-Fi), 802.16x (WiMedia), ZigBee, cell phones, Bluetooth and other relevant standards
• Interoperability of SP100 devices

Suppliers working on the standard have now coalesced into two groups, each advocating a similar field network, but differing approaches to control system integration. The WNSIA (Wireless Networks for Secure Industrial Applications) group includes Honeywell, Adaptive Instruments, 3eTechnologies International, Endress+Hauser, Flowserve, Omnex and Yokogawa. WNSIA proposes a field network centered about a set of field access points to reduce the number of “hops” necessary to get data from sensors to controllers. Its rationale is to minimize latency for feedback control connections. Additionally, WNSIA also proposes a fully integrated protocol linking the field network to legacy control systems through a gateway device.

The Collaborative Initiative (CI) group is centered about a very complete network architecture proposed by Dust Networks with backing from members Emerson, Invensys (Foxboro), Apprion, Certicom, Sensicast, Advanced Industrial Networks, Siemens, General Electric, Software Technology Group, NewTrax, Machine Talker, Oak Ridge National Laboratory and Texas Instruments/Chipcon. The CI/Dust proposal calls for a self-healing wireless mesh field network using channel-hopping. The mesh network provides greater distances at low power consumption and alternate routing in case of obstructions and interference sources.

Physical Layer and MAC
While the committee acknowledges that radio technology will change rapidly over the next few years, a complete standard must include at least one radio. The debate over the use of a licensed radio frequency for industrial automation has been shelved for two reasons:

• A licensed radio frequency band would require custom radios that would probably cost too much for this narrow market, and
• The time and effort to secure a single worldwide frequency for industrial automation use is both time-consuming and out of the scope of the SP100 committee.

The type of radio selected depends on the operating frequency (band) and is the subject of the Physical Layer subcommittee. The Media Access Control (MAC) sublayer is usually defined in conjunction with the Physical Layer, since it must provide the firmware interface for the radio.

Both groups have proposed networks based on the use of IEEE 802.15.4 (ZigBee) radios at 868/915 MHz or 2.4 GHz with frequency- (channel-) hopping on top of the base radio’s DSSS (Direct Sequence Spread Spectrum) protocol. These radios are designed to use very little power and be included on a battery-powered, solar cell or energy-scavenging device. Working groups are now meeting to decide on the specific physical layer and medium access sublayer protocols to be included in the standard. The committee is also looking at 802.15.4a radios recently approved by the IEEE 802.15.4 committee that use either of two forms of Ultrawideband (UWB) protocol:

• Chirp-OFDM (Orthogonal Frequency Division Multiplexing) operating in the 2.4 GHz band, or 
• Impulse Radio operating in the unlicensed UWB 3.5 to 10 GHz band.

These radios are expected to become very popular in many applications and will be very low- cost as well as have very low power consumption.

Using frequency or channel-hopping on top of the DSSS or OFDM protocols is the committee’s preferred method of avoiding interference from other radios such as Wi-Fi and ZigBee also operating in the unlicensed 2.4 GHz bands. Channel-hopping also helps to remove most of the adverse effects of multipath distortion often found in industrial locations. Not enough is known yet about practical use of impulse radio, but due to its very wide bandspread, it is also likely to be free of multipath distortion without the complexity of channel-hopping.

Use of these technologies for SP100 wireless field networks means that data transfer rates are expected to be low—on the order of 250 kbps over distances up to about 100 meters per radio hop. By using field-mounted access points (as in the WNSIA proposal) or mesh networks (as in the Collaborative Initiative proposal), much longer distances will be possible. The PHY-MAC (Physical Layer) Committee is now working on a compromise or union of the two proposals for the SP100 Physical Layer and MAC.
Upper Layers

At this point, very little work has been devoted to definition of the Data Link Layer (DLL) immediately above the MAC or any of the higher layers. Rather, considerable effort has been spent to study the plant environment into which SP100 networks must be integrated—a legacy environment. This information will be incorporated into a network and system architecture. Specifics of the SP100 architecture are not yet firm, but we have some data:

• SP100 is NOT being defined to be wireless Foundation fieldbus or to compete directly with wireless HART.
• It is likely that most of the defined IEC 61158 fieldbuses will need to be integrated with SP100. This may mean that fieldbus messages will need to be tunneled through an SP100 network, and that SP100 messages may need to be tunneled through any of the fieldbuses or routed to a fieldbus device.
• The decision to support Internet Protocol (IP) has not yet been finalized, but it now appears that IP addressing and protocol will not be directly supported for either IPv4 or IPv6.
• Work on the Application Layer (AL) has just begun. The committee has noted that the AL of Profibus and Foundation fieldbus are rarely, if ever, used because the DLL has been defined with sufficiently rich interfaces that the AL is not needed.
• Security is probably the most sensitive issue, since wireless exposes the industrial network to external forces that have never before been a problem in wired industrial networks.
• From the more than 40 use-case examples collected, it appears that applications will vary from simple data acquisition to some complex remote control.

Politics, Collaboration and Hidden Agendas
SP100 committee members are from many types of suppliers and a few end users. At present, it appears that no supplier is attempting to preempt committee work to foster their own agenda. The level of cooperation and collaboration is high. End-user members currently represent process manufacturing with little representation from discrete parts manufacturing. As a direct result, the current objectives of the main working group, SP100.11a is to define a wireless network architecture and protocol for typical continuous and batch process control applications. Each of the two groups has different and generally complementary technology to offer. The challenge will be to develop a single standard serving the end user with an inexpensive, highly reliable, secure, and easy-to-use wireless network.

Wireless HART is also progressing at a rapid pace with its own agenda to supplement the wired 4-20 ma HART specification with an appropriate wireless technology similar to that of SP100’s two groups. All groups use IEEE 802.15.4 radios in a similar way. No cooperation or interoperability discussions are currently under way or even planned between wireless HART and SP100. However, many of the same suppliers and committee experts are on both committees and offer some opportunity for cooperation.

Conclusions and Concerns
The SP100 committee plans to produce a standard by 2008 and currently appears to be on schedule. Some process control vendor products are already very close to the proposed physical layer, while others are not. However, all suppliers seem to be willing to modify current products to use whatever becomes the SP100 standard. Vendor infighting seems to not be impeding progress in the committee.
A major concern is that end users have not provided very much insight to the committee on potential use of wireless networks in their operations, beyond the obvious replacement of wires, once they become available at reasonable prices. Introduction of a disruptive technology, such as wireless, has changed every other industry when it has achieved a reasonable pricing level; e.g., home networks and telephony. SP100 should enable wireless applications for process automation well beyond simple data acquisition and control, not inhibit it.

Update on Recent SP100 Committee Decisions

THERE WILL be a “first release” of SP100.11a timed to be sooner, rather than complete. The architecture is supposed to be sufficient for a complete industrial automation network optimized for process control applications.

• The radio to be used is the one specified for IEEE 802.15.4, sometimes called the ZigBee radio. It will be used differently than ZigBee by using channel-hopping to improve multipath effects and to coexist with other ISM radios in the 2.4 GHz band.
• The first release will be suitable for 1-second process control loops, with a latency of 100 ms or more.

The next meeting will be in Phoenix, Feb. 12-15.

Dick Caro is principal of CMC Associates, is a recognized expert on fieldbus technology and a founding member and Chairman of the Fieldbus standards committees. You can reach him at [email protected].